An Improved Texture Feature Extraction Method Based on Radon Transform

Abstract

Criminal investigation image database retrieval is one of the key tasks in Forensic Science. Tire pattern is an important type of image data for crime scene investigation. It has been found that the rotation of tire pattern image has a great influence on retrieval precision. To relieve this problem, this paper proposes a new texture feature extraction method based on Radon transform, for more efficient tire pattern retrieval. More specifically, the proposed algorithm includes three steps. Firstly, in order to relieve the effects of image rotation on texture features, Radon transform is used to project the tire pattern image onto Radon domain. Then, Dual Tree Complex Wavelet Transform (DT-CWT) is applied to the coefficients in every direction of Radon domain. Finally, the mean value, variance and energy of every sub-band are extracted as the texture feature of the image. Compared with using Ridge let transform, the proposed method can better solve the problem of image rotation. Experimental results on a set of real-world tire patterns show that the proposed method is effective for tire pattern texture feature extraction and it outperforms other existing methods compared.